Development, reliability and validity of a new measure of overall health for pre-school children

Background: Few comprehensive systems are available for assessing and reporting the overall health of preschool children. Objectives: ( i) To develop a multi-dimension health status classification system (HSCS) to describe pre-school (PS) children 2.5-5 years of age; (ii) to report reliability and validity of the newly developed measure. Design: Existing systems (Health Utilities Index, Mark 2 and 3) were adapted for application to a pre-school population. The new system was tested for acceptability, validity and reliability. Participants: Three cohorts of children and their parents from Canada and Australia were utilized: Cohort 1 (MAC)-101 3-years old very low birthweight (VLBW, <1500 g) and 50 same age term children from Canada; Cohort 2 (AUS)-150 VLBW 3-years old from Australia; Cohort 3 (OMG)-222 3-years old with cerebral palsy (CP) from Ontario. Methods: Parental intra-rater reliability was evaluated by completion of the HSCS-PS Parent questionnaire (MAC) at the clinic visit and again 14 days later. Health professionals (MAC) completed the HSCS- PS Clinician questionnaire. Percent agreement and Kappa values were used to assess parent-clinician agreement. Concurrent validity was tested in two populations of VLBW children (MAC and AUS) and a reference group of term children (MAC) by exploring the relationships between dimensions of the HSCS- PS and well-recognized norm-referenced measures: the Bayley Scales of Infant Development (BSID- II), the Vineland Adaptive Behavior Scales (VABS) and the Stanford-Binet (SB). Construct validity was tested by comparing ratings on both the HSCS- PS and the Gross Motor Function classification system (GMFCS) using a population of pre-school children with CP. Analyses were done using x(2), ANOVA and correlations with tau-b statistic. Results: The HSCS- PS has 12 dimensions and 3-5 levels per dimension. Response rate for parental intra-rater reliability was 95%, with percent agreement ranging between 86 and 100%. Kappa values for various dimensions ranged from 0.38 to 1.00. Inter-rater reliability between parents and clinicians showed agreement ranging from 72 to 100%. Kappa values ranged from 0.30 to 1.00. Concurrent Validity: There was a statistically significant gradient between HSCS-PS Mobility levels and motor scale scores of the BSID- II and VABS. A significant gradient also occurred when comparing HSCS- PS cognition levels to psychometric scores on the BSID- II and SB, as well as HSCS- PS self-care levels compared to VABS Daily Living scores. Discriminative and Construct validity: Birthweight category was shown to be a significant determinant of proportion of children with multiple HSCS- PS dimensions affected. In addition, HSCS- PS dimension levels were congruent with GMFCS levels where expected: mobility had excellent correlation; self-care, dexterity, speech and cognitive dimensions had moderate correlations. Conclusions: The HSCS- PS is readily accepted, quick to complete, widely applicable and provides a multi-dimensional description of health status. Preliminary assessments of reliability and validity are promising. The HSCS- PS can discriminate across populations by birthweight and shows strong relationships with standardized psychometric measures in comparable domains. It can provide a summary profile of functional limitations in various populations of pre-school children in a consistent manner across programs and in different settings.